Newest Linux: Trick or Treat?

Editorial: Linux developers must focus on improvements.

Apparently unconcerned with the symbolism, the Linux community is marching toward its self-imposed Oct. 31 feature-freeze date for the newest operating system kernel. While what will be Version 2.6 has many laudable improvements, theres still plenty of post-Halloween maturation needed so that Linux stops frightening end users—and doesnt appear to be masquerading as an enterprise-ready platform.

Linux still gets a wary eye in the enterprise due to its restricted features and growing concern about its security. One key drawback, which will be addressed in 2.6, is support for setting file access permissions using access control lists. Every other major Unix version already does this.

In addition, Linux is still too dependent on local user lists. Universal LDAP support and systemwide LDAP integration, not scheduled for 2.6, need to be built in. On the server, this isnt a big deal, but deploying desktops without comprehensive support for a global user directory can be a nightmare. To alleviate management headaches for groups of Linux desktops, some system for pushing administrative policy and configuration changes to multiple users is a must.

Linux is well on its way in addressing scalability for large systems of eight CPUs or more as well as with plans to include a more feature-rich threading API in 2.6 to match what other Unix versions provide. But robust volume and storage management, as well as better event logging and log viewing, will have to wait for a later Linux.

On the security front, the Linux apple has lost its shine and now faces challenges that the open-source process once seemed nimble enough to evade. Needed after 2.6 will be tighter default installations and fewer programs that need to run as root. In addition, the Linux community should solicit security certification and adopt trusted operating system concepts, such as mandatory access controls, only the beginnings of which will be in 2.6.

If the Linux community wants to achieve its long-standing goal of acceptance of its operating system across the enterprise, its trick must be to turn its enormous developer energy into a treat of steady improvements.